Fixes / 02

Serial tests

One test step holds most of the job.

What the numbers show

test ≥ 3 min, ≥ 40% of the job, one step ≥ 80% of test time

A single test command running in one process, or one job running every suite, means the wait grows with every test added. Split the suite across workers or jobs. Every test still runs; it just runs side by side.

The change

Shard the suite across a job matrix
strategy:  matrix:    shard: [1, 2, 3, 4]steps:  - run: npx playwright test --shard=${{ matrix.shard }}/4
Or run workers inside the job
- run: pytest -n auto        # pytest-xdist- run: go test -p $(nproc) ./...- run: cargo nextest run

Runner options

Options for this case. Judged on setup effort, undo path and sizing. Full comparison
SetupOne line: runs-on: graphide
UndoA pull request restores runs-on
SizingWarm machine sized to the job, up to 32 vCPU
SetupCreate a larger runner, change its label
UndoChange the label back
SizingLarger sizes on paid plans
SetupProvision and patch your own pool
UndoChange the label back, retire machines
SizingWhatever you provision
SetupCheck each vendor's docs
UndoCheck each vendor's docs
SizingCheck each vendor's docs

Verify the verdict stays the same

  1. 01Total tests reported after the change equals the count before.
  2. 02No test is marked skipped that ran before.
  3. 03The job's conclusion matches on the same commit.
  4. 04Re-run three times to catch order-dependent tests.